Are firms (getting) ready for the corporate sustainability reporting directive?

In general, firms that are not subject to the SR mandates may publish sustainability information on a voluntary basis. For example, firms in industries such as oil and gas were among the first to use environmental disclosure to legitimise their operations to the society (Deegan and Gordon 1996). Similarly, firms use voluntary SR to respond to the pressure from key stakeholders such as customers, employees and investors (Fernandez-Feijoo et al. 2014). In contrast, reporting mandates importantly shape the SR processes and disclosure practices adopted by firms. The existing literature investigates different factors of SR which we classify into internal drivers, regulatory drivers and barriers to SR.

First, we review internal drivers of SR. On the one hand, they are based on the legitimacy theory, which argues that firms use (voluntary) SR to convince the society that they operate responsibly and thereby maintain their legitimacy (Cho et al. 2015). For example, firms in industries with a more pronounced negative impact on the environment and society are more likely to implement SR than firms in industries with lower impact (e.g., Elalfy et al. 2021; Fallan 2016). Larger and more profitable firms, which tend to have greater public visibility and more resources, are more likely to disclose sustainability information because they face greater social and political pressure (e.g., Brammer and Millington 2005; Gallo and Christensen 2011; Wickert et al. 2016). On the other hand, the internal drivers are also linked to the stakeholder theory, which argues that firms have responsibilities to stakeholder groups that can affect or are affected by the firm’s actions (Clarkson 1995). Fernandez-Feijoo et al. (2014) show that firms use SR to respond to the pressure from key stakeholders. In recent years, we have observed an increase in stakeholder demand for sustainability information, which can be linked to easy access to information on firms’ sustainability activities (e.g., via internet-based services and platforms) in combination with public discussions on environmental and societal issues generated by large firms (e.g., Andreou and Besharov 2022; Waugh and Oldfield 2022). However, stakeholder groups are very diverse and have different objectives (Ayuso et al. 2014). For example, financial stakeholders are primarily concerned with the impact of sustainability matters on a firm’s value creation, while consumers focus on the firm’s impact on the environment and society (Flower 2015).

Second, prior literature examines the role of regulatory factors on SR, which is often manifested in the quality of SR. Christensen et al. (2021) argue that voluntary and mandatory SR lack quality because the information disclosed is often repetitive, boilerplate, and not necessarily tailored to the reporting firm. But even achieving high-quality reporting through reporting mandates is not easy. It requires a transformation from a mere financial reporting to an integrated SR, which means that firms must invest in resources, overcome administrative challenges, and incur costs associated with such a new conceptualisation and operationalisation (Baumüller and Grbenic 2021; Baumüller and Sopp 2022; Nguyen 2020). The cost of transforming the reporting processes may be one of the main reasons for lower SR quality and the implementation of voluntary SR.

Prior literature highlights several weaknesses in the regulation of SR, including the NFRD. For example, Aureli et al. (2018) find that the lack of assurance reduces the credibility of SR and question the appropriateness of the reporting and assurance discretion provided to firms by the NFRD. Similarly, Vander Bauwhede and Van Cauwenberge (2022) conclude that mandatory SR should be accompanied by assurance to enhance the credibility of reports. Next, the NFRD’s concept of materiality seems insufficient for reporting relevant information. Raith (2023) finds that the concept is ambiguous and leads to a contest for materiality, with a focus on either business risks or impacts, shareholders or stakeholders, business or social case, but never on more than one of these determinants. Fiandrino et al. (2022) therefore call for a double materiality approach in reporting mandates. Slacik and Greiling (2020) investigate the implementation of the Global Reporting Initiative’s materiality concept and find that firms’ coverage and quality of the material aspects do not meet the requirements for relevant and transparent reporting. Finally, the comparability of sustainability information across firms is hampered when reporting mandates provide firms with flexibility and lack specific reporting standards. Steinhöfel et al. (2019) find significant discrepancies between the sustainability reports of German manufacturers falling under the scope of the NFRD and conclude that more harmonisation is needed to improve comparability of sustainability information and performance.

Third, prior studies reveal some reporting barriers by documenting the challenges of complying with SR regulation. These studies are mostly country-specific and investigate, for example, Spain (Criado-Jiménez et al. 2008; Larrinaga et al. 2002), Norway (Vormedal and Ruud 2009), Poland (Matuszak and Rozanska 2021) and the U.S. (Peters and Romi 2013). Generally, they find that firms are more likely to comply with SR requirements if their industry is considered environmentally sensitive, penalties for non-compliance are high, and regulatory enforcement and monitoring are adequate. On the other hand, Criado-Jimenez et al. (2008) show that enhanced regulation and enforcement may lead firms to adopt complex concealment strategies rather than truthful reporting.

In the last two decades, SR and its regulation in the EU has evolved substantially. Initially, the Commission Recommendation 2001/453/EC (EC 2001) and the Directive 2013/34/EU (EC 2013) encouraged the inclusion of sustainability information in firms’ annual reports. Subsequently, the passing of the NFRD in 2014 mandated large listed firms, banks and insurance firms (i.e., large public interest entities) with more than 500 employees to publish sustainability information from fiscal year 2017 onwards (first reports published in 2018). The NFRD aimed to improve the comparability of reporting on the multidimensional nature of corporate sustainability and firms’ application of sustainability policies, and pushed for more standardisation.

The EU responded to the 2030 Agenda for Sustainable Development (UN 2015) by stressing its commitment to achieving the 17 SDGs and the need for ambitious policies in line with the SDGs and their targets (European Council 2017). For example, the SR regulation supports EU firms to be socially responsible through the respect of gender equality, the promotion of energy efficiency and renewable energy, and with regard to working conditions in firms (contributing to SDG 5, 7 and 8, respectively).

Next, as part of the European Green Deal (EC 2019), the NFRD was reviewed and a new directive was announced to replace it in April 2021. The effectiveness of the NFRD was questioned as many firms did not publish material information on sustainability matters, and the comparability and reliability of published information was very limited. Finally, the CSRD, which introduces more comprehensive and detailed reporting, assurance requirement, and the application of the newly issued ESRS, was passed in November 2022 and came into force in January 2023. Firms will have to apply it from the 2024 fiscal year and the first sustainability reports will be published in 2025. The number of affected firms is expected to be significantly higher (approximately 49,000 firms) compared to firms subject to the NFRD (around 11,700 firms) (Deloitte 2022), but the scope will gradually expand as presented in Fig. 1.

Full size image

The CSRD will affect many firms and the implementation of the new reporting requirements will be challenging, particularly for firms with no previous SR experience. In the first years, firms’ preparation process may also suffer from regulatory uncertainty around the ESRS and their exact reporting requirements, as the standards have only been adopted in July 2023. In addition, the increase in assurance mandates for sustainability reports will pose challenges for assurance providers. Therefore, we find it relevant to examine the impacts of the CSRD’s announcement and passing on SR disclosure practices as the outcome of the preparation process.

We posit that following the announcement and passing of the CSRD, firms begin to prepare for the new reporting requirements, which should be observed in their SR disclosure practices.

First, we investigate whether firms not subject to the NFRD but subject to the CSRD increase their voluntary SR as part of their efforts to prepare for the CSRD. In line with legitimacy theory, firms may be motivated to engage in voluntary SR and convey the legitimacy of their business to society. The motivation may be particularly high in the recent years, as we have observed a high relevance of sustainability matters for governments and stakeholders. Therefore, firms may decide to implement SR and follow the latest trends in active communication of sustainability information (Zrnić et al. 2020). Relatedly, and in line with stakeholder theory, the implementation of voluntary SR may also be a firm’s response to increased stakeholder demand for accountability and transparency (Fernandez-Feijoo et al. 2014; Fiandrino and Tonelli 2021). Moreover, the CSRD itself includes advice for early preparation, as processing sustainability information, adjusting the reporting process and governance structures require additional effort and investment (Baumüller and Grbenic 2021). On the other hand, non-listed and smaller firms falling within the scope of the CSRD may lack the resources and expertise to overcome these challenges (Afolabi et al. 2023). Thus, it remains an open question whether voluntary SR changes prior to the CSRD’ application. We phrase the first hypothesis in an alternative form as follows:

Second, we investigate whether firms begin to implement the new reporting requirements before the CSRD applies. We examine the extent to which firms are already incorporating aspects of the CSRD’s requirements into their SR disclosure practices. Recent developments in SR and the growing interest in voluntary assurance (Vander Bauwhede and Van Cauwenberge 2022; Zrnić et al. 2020) may motivate firms with existing SR experience to begin implementing the new reporting requirements early. Moreover, compared to the NFRD, the new requirements demand more sustainability performance information to be processed and reported, which in turn requires the implementation of advanced reporting processes and governance structures (Baumüller and Grbenic 2021). Hence, we expect the NFRD firms to maintain their legitimacy in the society by making additional efforts in advance and preparing for the CSRD in a timely manner. As the CSRD itself gives firms clear guidance on how to analyse their organisations from a sustainability perspective, industry experts believe that the preparation process should start early (KPMG 2022). Such developments would imply that firms enhance their SR disclosure practices. We phrase our second hypothesis in an alternative form as follows:

We selected our sample from firms operating in Germany, which is the largest European economy in 2022 measured by gross domestic product (Eurostat and IMF 2023). Moreover, its strict reporting regulation enforcement by the Federal Financial Supervisory Authority suggests high future compliance with the CSRD, as it is with the NFRD, as non-compliance imposes high costs on firms (Loew and Braun 2019).

We focused on the population of non-financial firms and therefore excluded financial firms with WZ 2008 code K (Classification of Economic Activities, issue 2008, the German Federal Statistical Office 2008). First, we selected 60 firms that are not subject to the NFRD (i.e., non-NFRD firms) as follows. We obtained a list of listed firms with less than 500 employees (with financial statements available for 2022) from the Dafne database and randomly selected 30 firms. Next, we obtained a list of large or medium-sized private firms (with financial statements available for 2022) from the database and again randomly selected 30 firms.

Second, we took a sample of firms that are subject to the NFRD (i.e., NFRD firms). We obtained a list of listed firms with 500 or more employees² (with financial statements available for 2022) from the Dafne database and randomly selected 60 firms. For these 120 firms and for the fiscal years 2020–2022, we collected firm-level data on net income, total assets, property, plant and equipment (PPE), number of employees, and industry affiliation. As shown in Table 1, we cleaned the initial sample of 120 firms, so that each firm-year observation had all firm-level data and each NFRD firm had a published sustainability report. We also excluded firms (non-NFRD and NFRD) whose sustainability information was consolidated in the parent firm’s sustainability report. This left us with a final sample of 297 firm-years.

To construct the dependent variables, we retrieved firms’ sustainability reports and sustainability information from their websites.³ The base year was set to 2020, before the announcement of the CSRD. We expected that the announcement in April 2021 would have had little or no impact on SR for fiscal year 2020. The years 2021–2022 were set as the period after the CSRD’s announcement and passing.

To empirically test H1 on the increase in voluntary SR after the CSRD is announced and passed, we estimated the following logistic model on the sample of non-NFRD firms:

The dependent variable SustReport_it identified if firm i published a sustainability report⁴ in year t. The main variable of interest was Directive_t, which indicated the announcement and passing of the CSRD (i.e., equal to 1 for years 2021–2022 and 0 for the base year 2020).

Based on prior literature, our model controlled for firm characteristics associated with voluntary SR. Firms in industries with more pronounced negative impact on the environment and society are more likely to actively communicate their activities and contributions to a sustainable economy than other firms (e.g., Elalfy et al. 2021; Fallan 2016). Based on their share in pollution, waste production, water and energy consumption, we classified industries such as energy, transport, manufacturing, construction, technology, food and fashion retail as such (e.g., Wieser et al. 2021; Rad et al. 2022) and created the variable IndustryNeg_i. However, we could not estimate model (1) with this control variable on our sample because it perfectly predicted the outcome (i.e., all firms with voluntary SR were from industries with more negative impact). Larger and more profitable firms are more visible and have more resources to communicate sustainability activities to their stakeholders (e.g., Brammer and Millington 2005; Gallo and Christensen 2011; Wickert et al. 2016). Following Aluchna et al. (2023) and Cuomo et al. (2022), we measured Size_it as the natural logarithm of total assets, ROA_it as net income over total assets, and PPE_it as the ratio of tangible assets of firm i in year t. Definitions of all variables are provided in Appendix 1.

To test H2 on the enhanced SR disclosure practices after the CSRD’s announcement and passing, we estimated the following ordinal logistic model on the sample of NFRD firms:

SustReportPractices_it captured the extent of selected disclosure practices used by firm i in year t. The selected disclosure practices were based on the reporting requirements of the CSRD. We self-constructed a score that included the following four components, each worth one point: (1) publication of a sustainability report; (2) application of a double materiality approach;⁵ (3) voluntary assurance of the sustainability report (usually in accordance with ISAE 3000 (revised)⁶ or standards published by the German Institute of Public Auditors (i.e., IDW EPS 352 (08.2022), IDW EPS 990 (11.2022a), IDW EPS 991 (11.2022b)⁷); and (4) disclosure of sustainability information in a management report. Note that the first component was a prerequisite for scoring in any of the next three components. While the order in which these three components were scored did not matter, a higher total score always meant enhanced SR disclosure practices. The main variable of interest Directive_t, as in model (1), indicated the announcement and passing of the CSRD. The control variables were as in model (1).

We began our analysis by discussing and comparing the descriptive findings for the non-NFRD and NFRD firms (presented in Table 2). Table 3 showed the components of the SR disclosure score over time. The comparison of SR between the two groups and over time provided initial insights into the SR developments around the CSRD’s announcement and passing.

First, the descriptive statistics for non-NFRD firms depicted the current situation of voluntary SR, i.e. before the application of the CSRD. The mean of the SustReport was 9%, indicating that very few firms published voluntary sustainability reports during our sample period. As demonstrated by the first component of the SR disclosure score showed in Table 3, our sample contained four firms (five in 2022) with a voluntary report. On the other hand, 62% of the observations were from industries with a more pronounced negative impact on the environment and society, so one might have expected more voluntary SR. Note that all firms with voluntary sustainability reports were from these industries. Compared to the NFRD firms, the average non-NFRD firm was smaller, had negative profitability⁸ and a lower PPE ratio. In addition, non-NFRD firms had a much lower number of employees (based on mean and median) than NFRD firms, indicating a likely lack of capacity (resources) for voluntary sustainability reporting. Overall, we interpreted the finding that only one tenth of the firms that will be under the CSRD mandate in two or three years published voluntary sustainability reports as unsatisfactory.

Second, the descriptives of the total sample showed that the mean of the SustReport was 0.540 because it was dominated by the mandatory SR of the NFRD firms. While 57% of the total observations belonged to industries with more negative impact on the environment and society, the percentage was higher for non-NFRD firms. On average, NFRD firms were significantly larger (15.027 vs. 10.827), more profitable (0.044 vs. −0.043), had a higher PPE ratio (0.260 vs. 0.142) and had many more employees (32,607 vs. 268). These differences were consistent with the narrower scope of the NFRD compared with the CSRD. They also indicated that non-NFRD firms had fewer resources available for voluntary SR as part of the preparation for reporting under the CSRD.

Third, Table 2 also allowed us to compare the SR disclosure practices of firms that published sustainability reports (i.e., NFRD firms and non-NFRD firms with voluntary SR). The majority of observations (92%) belonged to NFRD firms, as non-NFRD firms rarely published voluntary reports. The mean of the SustReportPractices score was 2.070 and could be interpreted as “average extent of disclosure practices used”. However, the mean score for NFRD firms (2.130) was significantly higher than for non-NFRD firms (1.380). Specifically, as shown in Table 3, only one voluntary report applied the double materiality approach (in years 2021–2022) and obtained assurance (in years 2020–2022). In both groups of firms, more than half of the firms were from industries with more negative impact. Interestingly, all non-NFRD firms with voluntary SR belonged to these industries. Finally, untabulated descriptive statistics documented that the NFRD firms were on average larger (15.027 vs. 12.613), more profitable (0.044 vs. 0.025), had a higher PPE ratio (0.260 vs. 0.115) and had more employees (32,607 vs. 192) than non-NFRD firms with voluntary SR.

Finally, Table 3 allowed for a univariate analysis of the components of the SR disclosure. They were presented for each year and separately for non-NFRD firms with voluntary SR and NFRD firms. By default, all NFRD firms published sustainability reports in years 2020–2022 but only four non-NFRD firms (five in 2022) published voluntary reports (publication of a voluntary sustainability report was also captured by the SustReport variable). Because so few non-NFRD firms published voluntarily, we tested our H2 only on NFRD firms. More importantly, we found that an increasing number of NFRD firms applied the double materiality approach and had their sustainability reports assured. While the fourth component (i.e., sustainability information disclosed in the management report) declined slightly in 2021, the overall trend was upward.

Table 4 showed industry composition of all unique firms in the final sample as well as non-NFRD and NFRD firms separately. The main industries were similar for non-NFRD and NFRD firms, i.e., professional, scientific and technical activities (27% and 40%, respectively), manufacturing (18% and 30%, respectively), and wholesale and retail trade, repair of vehicles and motorcycles (22% and 6%, respectively). However, the non-NFRD firms came from a wider range of industries and the second largest industry was the wholesale and retail trade.

Panels A and B of Table 5 presented correlation matrices for the variables used in the regression models (1) and (2), respectively. As expected, Panel A showed that IndustryNeg and Size were significantly positively correlated with the dependent variable SustReport, which captured voluntary SR. Panel B showed that Directive, Size and PPE were significantly positively correlated with the SR disclosure practices, and IndustryNeg was negatively correlated with them. These univariate correlations were as expected and also indicated enhanced SR disclosure practices following the announcement and passing of the CSRD. As the independent variables in both panels were not highly correlated with each other, we concluded that multicollinearity was not an important concern.

First, we estimated model (1) to test whether the CSRD’s announcement and passing increase voluntary SR of non-NFRD firms. The results were reported in column 1 of Table 6. Contrary to H1, the variable Directive was not significantly associated with SustReport. Multivariate analysis supported our descriptive findings, suggesting that non-NFRD firms did not increase voluntary SR following the announcement and passing of the CSRD. This was contrary to our expectation that firms that will be mandated to report in a few years’ time would start voluntary reporting earlier in preparation for the CSRD. It was also inconsistent with voluntary SR being a means of responding to increasing stakeholder pressure and maintaining legitimacy in the society. While the regression results did not support H1, we caution that our small sample, with only five voluntary reporting firms, may have affected the finding.⁹ Among the control variables, Size was, as expected, significantly positively associated with SustReport. This could indicate that smaller firms lacked the resources to engage in voluntary SR as part of their preparation for the new reporting requirements. It would also be consistent with Afolabi et al.’s (2023) finding that smaller firms were less likely implement SR due to lack of resources, expertise and perceived benefits.

Taken together, our findings pointed to an unsatisfactory situation in which many non-NFRD firms will have to apply the reporting requirements of the CSRD in a few years’ time but are not yet preparing. They were also consistent with industry experts questioning the timeliness of these firms’ preparation (KPMG 2023). Our findings should remind policymakers, regulators and assurance providers that smaller firms may need additional support and resources to adequately prepare for the CSRD.

Second, we estimated the ordinal logistic regression model (2) to test whether the CSRD’s announcement and passing enhance SR disclosure practices of NFRD firms. The results were reported in column 2 of Table 6. Consistent with H2, the variable Directive was significantly and positively associated with the SR disclosure score, suggesting enhanced SR disclosure practices. Combining this result with our descriptive statistics (Table 3), we showed that NFRD firms most significantly enhanced the use of the double materiality approach after the CSRD’s announcement and passing, followed by the voluntary assurance of their reports. Overall, the results suggested that firms within the scope of the NFRD began to implement the new reporting requirements in their SR disclosure practices in a timely manner.

Looking at the regression coefficients of the control variables, we found that Size and PPE ratio were significantly positively associated with SR disclosure score. The variable IndustryNeg was negatively associated with SR disclosure score, suggesting that firms from industries with less negative impact were more likely to implement new reporting requirements. Such findings are consistent with Vander Bauwhede and Van Cauwenberge (2022) who showed that firm size and industry affiliation drive voluntary SR assurance.

Baumüller and Grbenic (2021) conceptually argue that the reporting requirements of the CSRD demand excessive resources and question whether the burden placed on the firms is acceptable. We empirically analysed firms’ preparation for the new reporting requirements through observable SR disclosure practices. Our findings suggested that additional effort and resources are particularly critical for non-NFRD firms with no prior experience in SR. Furthermore, the CSRD requires the application of the ESRS, which were adopted in July 2023. Due to their late adoption, firms in our sample period may have faced additional uncertainty about the precise reporting requirements and may have rationally decided to wait with preparation for the CSRD. While this may be true for NFRD firms that already published sustainability reports, it may be a particularly plausible alternative explanation for the poor implementation of voluntary SR by non-NFRD firms. Any additional administrative and financial resources required to comply with the ESRS may pose challenges for firms, irrespective of their current reporting status and the first year of application of the CSRD.